Right now, we always assume that the audio sources are split, then we use the split sources as the starting points for our pipeline. Perhaps we should instead allow for splitting to be part of the pipeline and assume that the only source is the one provided as the input file.
Right now, we always assume that the audio sources are split, then we use the split sources as the starting points for our pipeline. Perhaps we should instead allow for splitting to be part of the pipeline and assume that the only source is the one provided as the input file.